Goldman Sachs’ Chief Information Officer, Marco Argenti, revealed that the company is internally experimenting with generative AI tools to aid developers in generating and testing code. Argenti, who joined Goldman Sachs from Amazon in 2019, stated that the technology is currently in the proof of concept stage and not yet production-ready.

Despite pushback from some banking giants on the use of generative AI products, Goldman Sachs is exploring the technology’s potential, with software engineers already utilizing some assisted coding technology. Generative AI tools are products that produce human-like text or images in response to prompts, with ChatGPT, Bard by Google, and Stable Diffusion by Stability AI among the most popular examples.

Argenti refrained from disclosing the generative AI products being used by Goldman Sachs and did not specify which bank division is employing the technology. The utilization of ChatGPT and similar products has the potential to significantly disrupt the financial services industry, with AI potentially driving investment decisions and automating customer service functions.

Although Argenti acknowledged that it is still early for the development of generative AI, he compared its potential to that of the internet’s early days. While stressing that AI should not replace software developers, he stated that the technology can aid in increasing their productivity, with developers already able to generate up to 40% of their code automatically using generative AI. Furthermore, the software is being used for both code testing and generation, creating a dualism that allows the machine to test the work and vice versa.